CODEWITHSUNDEEP
scala
zorgbargle
zorgbargle

Reputation: 730

What is the correct way to initialise a trait's fields at construction

I'm relatively new to Scala, and am trying to code some basic object-relational mapping for a hobby project. I have classes representing my application domain objects, for example:

class Employee(val name:String)

I would like my ORM layer to record the database key with this employee, but this should not clutter the public Employee class. I thought I could use a trait for this:

trait DBEntity {val id:Int}

Within my ORM layer, when an employee is queried, I would return:

new Employee("Bob") with DBEntity {val id=5}

This will allow my ORM layer to later retrieve the id for an employee. My questions are:

1) Is this the best way to mix-in the trait? It boils down to creating an inline anonymous class, I would have preferred something like new Employee("Bob") with DBEntity(5), but traits can't have constructor parameters so this won't compile.

2) I am considering making Employee a case class to aid matching later on, would the new anonymous subclass returned by the ORM layer still be suitable for matching?

Upvotes: 1

Views: 98

Answers (1)

Kevin Wright
Kevin Wright

Reputation: 49685

This is certainly valid, but you might find it cleaner to have Employee subclass DBEntity. In this case you'd possibly want to make the id field a def in the trait and override it with a val in the Employee class.

Either way, pattern matching will work on the case class.

Upvotes: 1

Related Questions